Dense#
Dense (or fully connected) hidden layers are layers of neurons that connect to each node in the previous layer by a parameterized synapse. They perform a linear transformation on their input and are usually followed by an Activation layer. The majority of the trainable parameters in a standard feed forward neural network are contained within Dense hidden layers.
Parameters#
| # | Name | Default | Type | Description |
|---|---|---|---|---|
| 1 | neurons | int | The number of nodes in the layer. | |
| 2 | l2Penalty | 0.0 | float | The amount of L2 regularization applied to the weights. |
| 3 | bias | true | bool | Should the layer include a bias parameter? |
| 4 | weightInitializer | He | Initializer | The initializer of the weight parameter. |
| 5 | biasInitializer | Constant | Initializer | The initializer of the bias parameter. |
Example#
use Rubix\ML\NeuralNet\Layers\Dense;
use Rubix\ML\NeuralNet\Initializers\He;
use Rubix\ML\NeuralNet\Initializers\Constant;
$layer = new Dense(100, 1e-4, true, new He(), new Constant(0.0));